Some recent studies add to these fears, predicting sizable job displacement from numerous forms of automation and artificial intelligence in virtually all corners of the economy. But just as automation will alter industries differently, its effects will be much more intensive in some regional economies.
To estimate the potential effects of automation in those areas, Governing utilized definitions in a University of Oxford study assessing the automatability of individual occupations, then compared them with the Department of Labor’s most recent occupational employment estimates for the 100 largest U.S. metro areas. About 65 percent of Las Vegas area jobs were found to be susceptible to automation, the highest in any metro area. Much of that stems from the region’s large armies of servers, food preparers, cashiers and other occupations thought to be highly automatable. El Paso, Texas, and Cape Coral-Fort Myers, Fla., similarly employ many of these workers, and registered the next-highest shares of potential automatability.
Professors Carl Frey and Michael Osborne, who conducted the Oxford study, assigned a probability to each occupation by evaluating the extent to which its work activities require “creativity, social intelligence and perception, and manipulation.” Retail sales accounted for the single largest number of possible job displacements as a result of automation in most regions. The New York metro area, for instance, employs more than 500,000 retail salespersons and cashiers. Predominantly low-wage food service jobs are susceptible to drastic change as well, both in the United States and overseas. Robots will start delivering Domino’s pizza orders in Hamburg, Germany, this summer.
Regions with higher education levels should fare better. But the Brookings Institution’s Mark Muro points out that there’s more to it than that. Physical jobs that are more complex or personalized -- the kinds you won’t find on assembly lines -- may actually be less vulnerable to automation than routine office jobs. “Often, lower-skill but physical, personal or direct-caring occupations seem quite durable,” Muro says.
Middle-class, white-collar jobs, on the other hand, can be significantly liable to automation. A forthcoming report from Brookings reviews hundreds of U.S. occupations, finding use and knowledge of digital skills doubled between 2002 and 2016 and led to a wide array of jobs being digitized, including those of office clerks, customer service representatives and accounting workers. “The middle is where there will be some of the most disruption,” Muro says.
Some well-paying jobs in demand today aren’t off-limits from automation, either. A McKinsey Global Institute study concluded that some of the jobs most at risk involve data collecting and processing. Around a quarter of the activities of attorneys and physicians were deemed to be potentially automatable.
Large regions with jobs least susceptible to computerization, using the Oxford study’s definitions, are high-tech centers, such as San Jose-Sunnyvale-Santa Clara, Calif., and Durham-Chapel Hill, N.C. Other metro areas with highly educated workforces such as Washington, D.C., and Boston similarly appear to have fewer jobs vulnerable to displacement. Regional economies relying heavily on education and health care may be less prone to automation because jobs requiring a high degree of human interaction are thought to be among the most resilient.
(Larger markers represent regions more susceptible to automation based on a University of Oxford study. View an interactive map here.)
msa-automation.png
Of course, widespread automation won’t happen overnight. McKinsey projected that half the work activities across the economy today could be automated by 2055. An analysis by PricewaterhouseCoopers concluded that 38 percent of American jobs were at “high risk” of automation by the early 2030s. McKinsey studied prior cases of technological upheaval, finding that the time between initial commercial availability and peak adoption ranged between eight and 28 years.
The biggest unknown at this point is whether automation will eliminate more jobs than it creates. Automation itself isn’t new, and prior advances in technology and industrialization haven’t brought about higher overall unemployment over the long term. But a growing number of academics are concluding that automation this time around could, in fact, wield noticeably more harmful effects on the workforce. One highly cited paper by economists Daron Acemoglu and Pascual Restrepo forecasts lower overall employment resulting from the introduction of more robots into the workplace.
Other researchers, notably ones at the Economic Policy Institute, argue that automation has not led and will not lead to higher joblessness. Experts appear to be divided almost evenly on this question: A 2014 Pew Research Center survey of experts found 48 percent agreeing that automation, robots and artificial intelligence will displace more jobs than they create by 2025.
While many unknowns remain, it wouldn’t hurt for policymakers to start thinking about how to respond.
Some state workforce boards are looking at the issue. States already typically maintain labor market information divisions that project which occupations will be in demand in future years. Preparing farms and their workers for automation was the subject of a recent meeting of the California State Board of Food and Agriculture. While there aren’t yet many programs that specifically address automation, some states are engaged in activities that could help alleviate the impact of job losses. Apprenticeships are gaining a lot of attention and are expanding to health care, finance and other fields where they haven’t been common before. “The model is being modified and they’re really trying to ramp it up,” says Scott Sanders, executive director of the National Association of State Workforce Agencies.
For workers displaced by automation, community and technical colleges will play a crucial role in the pursuit of new careers. The federal government, however, has historically focused little on workforce training, spending much less than other wealthy nations do. “We don’t do training in America, we do education,” says Anthony Carnevale, who directs the Georgetown University Center on Education and the Workforce. “Our policy is: Go to college.”
It was only a few short decades ago that computers began revolutionizing the American workplace. Regions and employers that were early adopters with skilled workforces are well ahead today, and it’s likely they will continue to be in the years to come.
